Bigram English Fitness Score

Score actual overlapping English bigrams against explicit independently supplied log corpus model/unknown penalty, no fabricated language corpus or decryption correctness.

Description

Bigram fitness sums corpus-dependent log scores for adjacent symbol pairs.

Bigram fitness sums corpus-dependent log scores for adjacent symbol pairs.^[1]

An illustrative caller corpus does not install authoritative English probabilities or prove a successful decryption. Use Bigram English Fitness Score to evaluate the supplied inputs under the specific model below.

When to use Bigram English Fitness Score

Use Bigram English Fitness Score to evaluate the supplied inputs under the specific model below. Choose the intended encoding, alphabet or check and keep the exact character, key and measurement conventions described below.

How to use Bigram English Fitness Score and understand its results

Choose the intended encoding, alphabet or check and keep the exact character, key and measurement conventions described below.
Enter Text, Caller bigrams, Caller log probabilities, Caller unknown log probability using the field table and validation rules below, then run the tool.
Read Total bigram log fitness, Scored bigram count, Unknown bigram count and interpret units, signs, ordering and placeholders as explained below.
If the tool reports an input error, correct that field and rerun. A successful false result, empty match list or unknown-evidence placeholder has the specific meaning described here; check the limitations before applying the result.

Input field	Meaning	Value type	Requirement
text	Text	string	Required; constraints below
callerBigrams	Caller bigrams	array of string	Required; constraints below
callerLogProbabilities	Caller log probabilities	array of number	Required; constraints below
callerUnknownLogProbability	Caller unknown log probability	number	Required; constraints below

Enter uppercase A–Z text, unique uppercase two-letter bigram keys, aligned finite nonpositive log probabilities and a nonpositive unknown-bigram fallback. Sum scores for every overlapping adjacent pair and count unknown pairs. Empty and single-letter text have no pairs and score zero. The supplied table need not form a normalized corpus and the tool has no built-in English model.

score = ΣoverlappingPairs suppliedLogProbabilityOrFallback; unknownCount = number of fallback pairs

Field order, units and branch conditions follow the implementation explanation above.

Output field	Meaning	Value type
totalBigramLogFitness	Total bigram log fitness	number
scoredBigramCount	Scored bigram count	number
unknownBigramCount	Unknown bigram count	number

json

{
  "input": {
    "text": "ABAB",
    "callerBigrams": [
      "AB",
      "BA"
    ],
    "callerLogProbabilities": [
      -1,
      -2
    ],
    "callerUnknownLogProbability": -5
  },
  "output": {
    "totalBigramLogFitness": -4,
    "scoredBigramCount": 3,
    "unknownBigramCount": 0
  }
}

Limitations and assumptions

An illustrative caller corpus does not install authoritative English probabilities or prove a successful decryption. Apply the exact scope and assumptions described above.

All listed fields require explicit values. Unexpected fields, incorrect types, nonfinite numbers and sparse arrays fail validation. Domain-specific bounds, alignments and supported tokens are stated above; extremely large inputs or outputs can also be rejected. These errors differ from successful results such as false, no matches or unknown evidence.

Alternative or Complementary methods

Use Chi-Squared English Fitness Scorer for a related task: A Pearson letter-fit score compares observed counts with a supplied positive frequency model.

Use Vigenère Key-Length IoC Estimator for a related task: Index of coincidence compares repeated letters within candidate key-position columns.

References

Bigram English Fitness Score — tool.io

Similar or alternative tools

Chi-Squared English Fitness Scorer
A Pearson letter-fit score compares observed counts with a supplied positive frequency model.
Vigenère Key-Length IoC Estimator
Index of coincidence compares repeated letters within candidate key-position columns.